KITE (Knowledge-Informed Tutoring Engine)

Product or Programme · All products and programmes
Evidence · 4 claims (2 for, 2 mixed) · 2 studies (2 design), q2 · 0 of 2 report an effect size · 4 claims rest on one study

Description

KITE is a course-grounded intelligent tutoring system developed by Mragisha Jain et al. that combines multi-stage retrieval with intent-routed pedagogical response strategies for algorithmic reasoning instruction.

Components

• KITE: an intent-aware RAG tutoring architecture combining multi-stage retrieval with pedagogical response strategies: KITE (Knowledge-Informed Tutoring Engine) is a RAG-based intelligent tutoring system for course-grounded dialogue on algorithmic reasoning and problem-solving. As the article states, "the system includes five components: document preprocessing, embedding generation, multi-stage retrieval, intent-aware response generation, and session management." Each query is classified by pedagogical intent, and the classified intent determines the response strategy: grounded explanations for direct and conceptual questions, Socratic assessment for validation, diagnostic hint progressions for debugging, and step-by-step state-tracing for algorithm tracing. Retrieval combines dense bi-encoder search, hybrid scoring (70% dense, 30% BM25), MMR with λ = 0.7, cross-encoder reranking, and source-based boosting, passing the top eight chunks to GPT-5. (Mragisha Jain et al. (2026))

Claims

• KITE's non-procedural responses show strong retrieval grounding (faithfulness 0.85, context relevance 0.94) and answer similarity 0.76, with lower factual correctness 0.45 [+M]
• KITE's feedback led simulated students to improve revised answers in 24 of 27 non-already-correct interactions (88.89%) [+M]
• Evidence tracing links 96 of 114 generated claims to source chunks, but three claims with retrieval scores of 0.50 or below show retrieval similarity alone is insufficient [~W]
• Research-question responses score higher Answer Relevancy (0.87) than follow-ups (0.72), while Faithfulness, Context Relevance and Response Groundedness are lower [~W]

Related Products and Programmes

-

Key Sources

• Mragisha Jain, Tirth Bhatt, Griffin Pitts, Aum Pandya, Peter Brusilovsky, Narges Norouzi, Arto Hellas, Juho Leinonen, Bita Akram. (2026). Retrieval-Augmented Tutoring for Algorithm Tracing and Problem-Solving in AI Education. Proceedings of the 21st Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2026), co-located with ACL 2026. https://arxiv.org/abs/2605.12988

Page metadata

Field	Value
type	product
id	kite-knowledge-informed-tutoring-engine
title	KITE (Knowledge-Informed Tutoring Engine)
description	KITE is a course-grounded intelligent tutoring system developed by Mragisha Jain et al.
status	review
product_kind	software
generated	by process:settle-candidates on 2026-10-11
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.